Draft version December 31, 2018Typeset using LATEX twocolumn style in AASTeX62

Indication of Another Intermediate-mass Black Hole in the Galactic Center

Shunya Takekawa,1 Tomoharu Oka,2, 3 Yuhei Iwata,3 Shiho Tsujimoto,3 and Mariko Nomura4

1Nobeyama Radio Observatory, National Astronomical Observatory of Japan

462-2 Nobeyama, Minamimaki, Minamisaku-gun, Nagano 384-1305, Japan2Department of Physics, Institute of Science and Technology, Keio University

3-14-1 Hiyoshi, Kohoku-ku, Yokohama, Kanagawa 223-8522, Japan3School of Fundamental Science and Technology, Graduate School of Science and Technology, Keio University

3-14-1 Hiyoshi, Kohoku-ku, Yokohama, Kanagawa 223-8522, Japan4Astronomical Institute, Tohoku University

6-3 Aramaki, Aoba, Sendai 980-8578, Japan

Submitted to ApJL

ABSTRACT

We report the discovery of molecular gas streams orbiting around an invisible massive object in the

central region of our Galaxy, based on the high-resolution molecular line observations with the Atacama

Large Millimeter/submillimeter Array (ALMA). The morphology and kinematics of these streams can

be reproduced well through two Keplerian orbits around a single point mass of (3.2±0.6)×104 M. We

also found ionized gas toward the inner part of the orbiting gas, indicating dissociative shock and/or

photoionization. Our results provide new circumstantial evidences for a wandering intermediate-mass

black hole in the Galactic center, suggesting also that high-velocity compact clouds can be probes of

quiescent black holes abound in our Galaxy.

Keywords: Galaxy: center — ISM: clouds — ISM: molecules — submillimeter: ISM

1. INTRODUCTION

Intermediate-mass black holes (IMBHs) with masses

of 102–105 M are the missing link between stellar-mass

and supermassive black holes (e.g., Ebisuzaki et al. 2001).

Many efforts have been made to confirm the exis-

tence of IMBHs (Mezcua 2017). Ultra-luminous X-

ray sources have been considered as promising IMBH

candidates (e.g., Farrell et al. 2009). Relatively mas-

sive IMBHs may lurk in the nuclei of dwarf galax-

ies and/or globular clusters (e.g., Reines et al. 2013;

Baldassare et al. 2015; Kızıltan et al. 2017). However,

these results have been argued upon, and therefore,

none of the IMBH candidates are accepted as definitive

(e.g., Ebisawa et al. 2003; Strader et al. 2012).

High-velocity(-width) compact clouds (HVCCs),

which are a population of compact molecular clouds

with extremely broad velocity widths (Oka et al. 1998;

Oka et al. 1999), may also provide possible hints of

Corresponding author: Shunya Takekawa

shunya.takekawa@nao.ac.jp

the existence of IMBHs. HVCC CO–0.40–0.22 has

been interpreted as a cloud that was gravitationally

kicked by a massive IMBH with a mass of 105 M(Oka et al. 2016; Oka et al. 2017). Although the puta-

tive IMBH has a radio emission counterpart CO–0.40–

0.22∗, the nature of CO–0.40–0.22 and CO–0.40–0.22∗

is still controversial (Ravi et al. 2018; Tanaka 2018).

IRS13E, which is an infrared source in the vicinity of

the Galactic nucleus Sgr A∗, has also been suggested to

contain a ∼ 104 M IMBH based on the high-velocity

feature of the ionized gas (Tsuboi et al. 2017) as well

as the stellar dynamics (Schodel et al. 2005).

Recently, we discovered a peculiar HVCC, HCN–

0.009–0.044, at a projected distance of 7 pc from Sgr A∗,

by using the James Clerk Maxwell Telescope (JCMT)

(Takekawa et al. 2017). HCN–0.009–0.044 is more com-

pact (∼ 1 pc) than any previously known HVCCs (2–5

pc), and its velocity width (∼ 40 km s−1) is typical to

those of HVCCs. The compactness, kinematics, and ab-

sence of luminous stellar counterpart can be explained

by the high-velocity plunge of an invisible compact

object into a molecular cloud (Takekawa et al. 2017;

arX

iv:1

812.

1073

3v1

[as

tro-

ph.G

A]

27

Dec

201

8

2 Takekawa et al.

Nomura et al. 2018). The driving source may be an in-

active and isolated black hole.

This Letter reports on the results of high-resolution

observations of HCN–0.009–0.044 with the Atacama

Large Millimeter/submillimeter Array (ALMA) and the

discovery of molecular gas streams showing clear orbital

motions around an invisible gravitational source with

a mass of (3.2 ± 0.6) × 104 M. The distance to the

Galactic center is assumed to be D = 8 kpc.

2. OBSERVATIONS

Our ALMA cycle 5 observations (2017.1.01557.S) were

performed on May 13–14, 2018. Eleven 7-m antennas

and forty-six 12-m antennas were used to obtain the

HCN J=4–3 (354.5 GHz), HCO+ J=4–3 (353.6 GHz),

and CS J=7–6 (342.9 GHz) datasets of the target source

HCN–0.009–0.044. The field of view was 54′′× 54′′ cen-

tered at (l, b) = (−0.009, −0.044), which was covered

with 7 and 23 pointings of the 7-m and 12-m arrays, re-

spectively. The on-source times were 43.34 and 9.27 min

for the 7-m and 12-m array observations, respectively.

The 12-m array configuration was C43-2 with baseline

lengths of 15–314 m. The bandwidths of the spectral

windows for the HCN, HCO+, and CS observations were

respectively 1, 0.5, and 2 GHz with 1.953-MHz channel

widths. J1924–2914 and J1517–2422 were observed as

the flux and bandpass calibrators. The phase calibrators

were J1744–3116 and J1700–2610.

We calibrated and reduced the data by using the Com-

mon Astronomy Software Applications (CASA) software

package (version 5.1.2-4) in the standard manner1. The

calibration was performed with the calibration script

provided by the East Asian ALMA Regional Center.

The visibility data were split into spectral lines and con-

tinuum emissions through the task ‘uvcontsub’. The

synthesized images were created using the task ‘tclean’

with Briggs weighting with a robust parameter of 0.5.

The spatial and velocity grid width of the resultant im-

age cubes were 0.5′′ and 2 km s−1, respectively. The

synthesized beam sizes were as follows: 0.87′′ × 0.71′′

with a position angle (PA) of −31.6 at 354.5 GHz,

0.87′′ × 0.71′′ with a PA of −30.3 at 353.6 GHz, and

0.89′′ × 0.73′′ with a PA of −30.7 at 342.9 GHz. The

root mean square (rms) noise level of the image cubes

was 7 mJy beam−1.

3. RESULTS AND DISCUSSION

3.1. Spatial/Velocity Structure and Physical Condition

Our high-resolution observations have unveiled the en-

tity of HCN–0.009–0.044. Figure 1(a) shows the inte-

1 https://casaguides.nrao.edu/index.php/ALMAguides

grated intensity map of the HCN J=4–3 line. HCN–

0.009–0.044 is resolved into several components. The

main body appears as a balloon-like structure (hereafter

Balloon). A stream-like structure (hereafter Stream)

lies on the southeast of the Balloon. An ultra-compact

clump (UCC) is associated with the tip of the Stream.

Figure 1(b) shows the averaged line-of-sight velocity

(moment 1) map of the HCN line. The Balloon and

Stream clearly exhibit velocity gradients. The gas of

the Balloon is gradually accelerated clockwise from the

northeastern part at the line-of-sight velocity of VLSR ∼−70 km s−1, indicating a rotational motion (see also Fig-

ure 2). The Stream is accelerated from south to north (

VLSR ∼ −60 km s−1 to −40 km s−1) as if it is orbiting

around the center of the Balloon.

HCN–0.009–0.044 is likely to be at least 5 kpc apart

from the Sun (R > 5 kpc) because CO J=3–2 emis-

sion from HCN–0.009–0.044 (Parsons et al. 2018) suf-

fers from absorption by molecular gas in the 3-kpc arm,

which is the Galactic spiral arm at ∼ 5 kpc (Sofue 2006).

The components traced by the HCN J=4–3 emission

also appear in the ALMA images of the HCO+ J=4–

3 and CS J=7–6 lines (Figures 1(c) and (d)). HCN–

0.009–0.044 consists of highly dense and hot molecular

gases as the critical densities for these transitions are

∼ 107 cm−3, and the J=7 level of CS is 65.8 K above

the ground state. Such dense/hot molecular gases are

abundant in the central 200 pc of our Galaxy but rare

in the Galactic disk region. Hence, HCN–0.009–0.044

is probably in the Galactic center (R = 8 kpc). The

molecular gas masses of the Balloon, Stream and UCC

were estimated to be MLTE ∼ 6 M, 1 M and 0.1

M, respectively, assuming the local thermodynamic

equilibrium (LTE) with an excitation temperature of

22 K in the optically thin limit (Takekawa et al. 2017)

and a fractional abundance of [HCN]/[H2] = 4.8× 10−8

(Tanaka et al. 2009; Oka et al. 2011).

Figure 1(e) shows the continuum image at 354.5 GHz.

A faint filamentary structure appears in the eastern side

of the Balloon. Although this filament also appears in

the 353.6- and 342.9-GHz bands, we could not derive the

reliable spectral index because of the insufficient sensi-

tivities. M–0.02–0.07 (also referred to as the +50 km

s−1 cloud) and the Northern Ridge at VLSR ' 0 km s−1

(McGary et al. 2001) overlap with the filament in the

line of sight. The filament may reflect dust emission

from them. The physical relation between the contin-

uum filament and the HVCC is currently ambiguous.

3.2. Origin of HCN–0.009–0.044

Interactions with supernova remnants (Oka et al. 2008;

Tanaka et al. 2009; Yalinewich & Beniamini 2018) and

Indication of Another IMBH in the Galactic Center 3

Ga

lactic L

atitu

de

[d

eg

]–

0.0

45

–0

.05

–0

.04

0 3 6 9 12 –75 –60 –45 –30

[ km s–1 ]

HPBW HPBW

[ Jy beam–1 km s–1 ]

UCC

Stream

UCC

Stream

(a) (b)

Balloon Balloon

Galactic Longitude [deg]

Ga

lactic L

atitu

de

[d

eg

]

–0.005 –0.015–0.01

–0

.04

5–

0.0

5–

0.0

4

HCN J=4–3

Galactic Longitude [deg]

–0.005 –0.015–0.01

0 2 4 6

HPBW

[ Jy beam–1 km s–1 ]

UCC

Stream

(c)

Balloon

HCO+ J=4–3

Velocity field

0 2 4 6

HPBW

[ Jy beam–1 km s–1 ]

UCC

Stream

(d)

Balloon

CS J=7–6

Galactic Longitude [deg]

–0.005 –0.015–0.01

0 2 4 6

HPBW

[ mJy beam–1 ]

(e)Continuum

(354.5 GHz)

8

0.2 pc 0.2 pc

0.2 pc 0.2 pc 0.2 pc

Figure 1. (a) Integrated intensity map of the HCN J=4–3 line over VLSR = −80 – −20 km s−1. The gray contours show thesame map obtained by using the JCMT (Takekawa et al. 2017). The magenta ellipse indicates the extent of the Balloon. Thehalf-power beam widths (HPBWs) of ALMA and JCMT are represented by a white filled ellipse and gray circle, respectively.(b) Averaged line-of-sight velocity (moment 1) map of the HCN line. (c, d) Integrated intensity maps of the HCO+ J=4–3 andCS J=7–6 lines over VLSR = −80 – −20 km s−1. (e) Continuum image at 354.5 GHz obtained with the 1-GHz bandwidth. Therms noise level of the continuum is 2 mJy beam−1.

cloud–cloud collisions (Tanaka et al. 2015; Tanaka 2018;

Ravi et al. 2018) have been considered as the origins of

HVCCs. However, the position–velocity structure of

HCN–0.009–0.044 exhibits neither expanding motion

indicative of a supernova–cloud interaction (e.g., see

Figure 12 in Fukui et al. 2012) nor V-shaped pattern

characteristic of a cloud–cloud collision (e.g., see Figure

10(c) in Torii et al. 2017). Natural explanations for the

velocity gradients along the Balloon and Stream (Figure

1(b)) are orbital motions caused by gravity. The ob-

served kinematics implies that a massive gravitational

source lurks in the Balloon. The detection of the CS

J=7–6 line (Figure 1(d)), which is a high-density and

possibly shock probe (e.g., Tanaka et al. 2018), may

support the interpretation that the molecular gas has

experienced the tidal compression by the strong gravity.

The enclosed mass (M) can be roughly estimated by

M ∼ ∆V 2R∆θ/2G, where ∆V is the velocity width, R

is the distance to the cloud, ∆θ is the angular diameter,

and G is the gravitational constant. The observed veloc-

ity width and diameter of the Balloon are ∆V ∼ 20 km

s−1 and R∆θ ∼ 0.4 pc, respectively. Thus, a massive

object with ∼ 104 M may be hidden in the Balloon.

3.3. Keplerian Model

In order to confirm whether Keplerian motions explain

the kinematical structures of the Balloon and Stream,

we performed orbital fittings to the observed data with

the similar procedure as done for the Sgr A West by

Zhao et al. (2009). Assuming the dynamical center at

(l, b, R) = (−0.0096, −0.0451, 8 kpc), we deter-

mined three-dimensional orbital parameters (a, e, Ω, ω,

i) for the orbital geometries of the Balloon and Stream

by least-square fitting to the loci we chose (red and blue

points in Figure 3(b)). We carefully chose these loci

(including the dynamical center) so that they represent

the kinematics of the observed features. The orbital

parameters are the semi-major axis (a), eccentricity (e),

longitude of ascending node (Ω), argument of pericen-

ter (ω), and inclination angle (i). We set the X-, Y-,

and Z-axes parallel to the Galactic longitude, Galactic

latitude, and line-of-sight direction, respectively. The

best-fit orbital parameters are listed in Table 1 and the

4 Takekawa et al.

–0.005 –0.01 –0.015

–0.045

–0.05

–0.04

Galactic Longitude [deg]

Ga

lactic L

atitu

de

[d

eg

]

[ Jy beam–1 km s–1 ]0 2 4 6

HPBW

Stream Stream

UCC

Stream

UCC UCC UCC UCC

0.2 pc

0.2 pc

0.2 pc

0.2 pc 0.2 pc

0.2 pc0.2 pc

0.2 pc

Figure 2. Velocity channel maps of HCN–0.009–0.044. Each panel shows the HCN J=4–3 emission integrated over 10 km s−1.The numerical value on the upper left corner in each panel indicates the central velocity in km s−1. The white filled ellipse onthe lower left corner indicates the HPBW of ALMA. The white contours show the Pα emission at 1.87 µm (Dong et al. 2011).The contour levels are 300, 350, and 400 µJy arcsec−2. The rms noise level of the Pα map is ∼ 74 µJy arcsec−2.

Table 1. Parameters of the modeled Keplerian orbits

Parameters Balloon Stream

Semi-major axis, a 0.20 ± 0.02 pc 0.54 ± 0.15 pc

Eccentricity, e 0.66 ± 0.05 0.62 ± 0.31

Longitude of ascending node, Ω 62 ± 24 deg −113 ± 41 deg

Argument of pericenter, ω −99 ± 21 deg −115 ± 34 deg

Inclinationa, i 153 (or 27) ± 13 deg 136 (or 44) ± 21 deg

Pericenter distance 0.07 pc 0.21 pc

Orbital period 4.6 × 104 years 21.1 × 104 years

a If the orbital motion is counterclockwise, the inclination of i < 90 is chosen.

resultant orbits are projected in Figures 3(a)–(c). Note

that there remains a double degeneracy in the inclina-

tion angle. Orbits with i and (180 − i) produce the

same projected orbits and line-of-sight velocities.

The line-of-sight velocity (VZ) at each (X, Y) of the

orbits depend on the mass (Mdyn) and line-of-sight ve-

locity (Vdyn) of the dynamical center. After determining

the orbital geometries, we fitted the modeled velocities

VZ to the observed velocities on the orbits of the Bal-

loon and Stream with free parameters of Mdyn and Vdynthrough a chi-square (χ2) minimization approach. We

used the moment-1 values of the HCN image smoothed

with a Gaussian kernel of 3′′ full width at half maxi-

mum (Figure 3(a)) for the model fit to reduce the noise

effect. As a result, we derived the best-fit values of

Mdyn = (3.2 ± 0.6) × 104 M and Vdyn = −49.5+1.0−0.7

km s−1. Figure 3(c) shows the velocity field of the

model with the best-fit parameters. Figure 3(d) shows

the confidence level contours as a function of Mdyn and

Vdyn. These suggest that the observed kinematics of the

Balloon and Stream can be well explained by two Ke-

plerian orbits around a single gravitational source with

Mdyn ' 3× 104 M.

Figure 4 shows the three-dimensional schematic view

of HCN–0.009–0.044 based on the model with the best-

fit parameters. The cloud lying on the south of the Bal-

loon with VLSR ∼ −60 km s−1 may share the same orbit

as the Stream. The dynamical time scales of the Bal-

loon and Stream are estimated to be 4×104 and 6×104

years, respectively. The similarity of the dynamical time

scales may indicate that the Balloon and Stream are de-

rived from a common parent molecular cloud. These

Indication of Another IMBH in the Galactic Center 5

Figure 3. (a) Averaged line-of-sight velocity (moment 1) map of the HCN line smoothed with a 3′′ Gaussian kernel. The whitefilled circle on the lower left corner indicates the full width at half maximum of the Gaussian kernel. The magenta contoursshow the Pα emission (Dong et al. 2011). (b) Loci used for the orbital fittings for the Balloon (red) and Stream (blue). Thebackground image (grayscale) is the HCN integrated intensity map. The ellipses are the modeled orbits for the Balloon andStream, respectively. Their orbit parameters are listed in Table 1. The star indicates the dynamical center. (c) Color mapof the line-of-sight velocities (VZ) of the modeled orbits with the best-fit values. The contours show the HCN intensity of 1Jy beam−1 km s−1. The ellipses and star are the same as those in the panel (b). (d) Confidence level contours as a function ofMdyn and Vdyn. The contour levels are 1σ (68.3%), 2σ (95.4%), and 3σ (99.7%). The cross mark indicates the best-fit values.

6 Takekawa et al.

–0.4–0.2

0 0.2

0.4 0.6

–0.8

–0.6

–0.4

–0.2

0

0.2

0.4

–0.6

–0.4

–0.2

0

0.2

0.4

X [pc]

Y [pc]

Z [pc]

01

23

45

6 78

10

11

00.5 1.0 1.5

2.02.5

3.03.54.0

9

Line of sight

Figure 4. Three-dimensional schematic view of HCN–0.009–0.044 based on the Keplerian model with the best-fitparameters. We adopt clockwise motions (i > 90) for theorbits. The red and blue arrows indicate the direction of theorbital motions for the Balloon and Stream, respectively, Thered, green, blue and orange clouds represent the Balloon,Stream, −60 km s−1 clump, and UCC, respectively. Theblack circle indicates the position of the IMBH. The red andblue points on the orbits, which are also projected on theX–Y plane, indicate the elapsed times. The red and bluenumerical values are the elapsed times in unit of 104 yr. Thegrayscale on the X–Y plane is the HCN map.

results are consistent with the notion that the Balloon

and Stream have been captured by the gravitational po-

tential well of a massive compact object in the relatively

recent past.

It should be noted that the UCC shows an abnormally

broad velocity width which can not be explained by the

Keplerian model. The UCC is located at the northern

tip of the Stream, having velocities from VLSR ' −50

km s−1 to 0 km s−1 without velocity gradient (Figure 2).

The positive velocity end is contaminated with emission

from the Northern Ridge (McGary et al. 2001). The

virial mass is roughly estimated to be ∼ 104 M, which

is far greater than the LTE mass (∼ 0.1 M). The UCC

may contain a stellar object as a core, and the broad ve-

locity width could be attributed to an outflow from it.

Further observations with higher spatial resolution are

necessary to reveal the origin of the UCC.

3.4. Indication of an IMBH

According to the model, a huge mass of (3.2±0.6)×104

M must be concentrated within a radius significantly

smaller than 0.07 pc (the pericenter distance for the Bal-

loon’s orbit). The averaged mass density is much larger

than that of the core of M15 (ρ ∼ 2 × 107 M pc−3),

which is one of the most densely packed globular clus-

ters (Djorgovski & King 1984). However, considering

the star cluster as the gravitational source is implausible

because of the absence of luminous stellar counterparts.

Therefore, the most promising candidate for the gravi-

tational source is a massive IMBH. Although accreting

black holes can be traced through X-ray emission from

their accretion disks (e.g., Done 2010), no X-ray point

source has been detected in the extent of the Balloon

(Muno et al. 2009). This is possibly because the IMBH

is isolated and inactive. HCN–0.009–0.044 is the third

case of the possible IMBH holder in the Galactic center

after IRS13E (Schodel et al. 2005; Tsuboi et al. 2017)

and CO–0.40–0.22 (Oka et al. 2016; Oka et al. 2017).

Many black holes follow the Fundamental Plane of

black hole activity, which is an empirical relation be-

tween radio luminosity, X-ray luminosity and black hole

mass (e.g., Merloni et al. 2003; Falcke et al. 2004). Dif-

fuse continuum emission at 5.5 GHz (Zhao et al. 2013;

Zhao et al. 2016) overlaps with HCN–0.009–0.044 (see

Figure 2(b) in Takekawa et al. 2017). This sensitive

image provides an upper limit radio luminosity of ∼2× 1028 erg s−1 for the putative IMBH. If the IMBH is

placed on the Fundamental Plane derived by Plotkin

et al. (2012), then the X-ray luminosity can be in-

ferred to be . 1031 erg s−1 from the radio luminos-

ity of . 2 × 1028 erg s−1 and the mass of 3.2 × 104

M. On the other hand, the X-ray image at 0.5–8 keV

(Muno et al. 2009) provides an upper limit X-ray lumi-

nosity of ∼ 7×1030 erg s−1 for the IMBH. This is consis-

tent with the Fundamental Plane for low accretion rate

black holes (Plotkin et al. 2012), thereby providing an-

other support for the IMBH interpretation.

Although there is no point source suggestive of the

IMBH, we found an emission counterpart for the Bal-

loon in the Pα recombination line at 1.87 µm obtained

through the Hubble Space Telescope (Dong et al. 2011)

(Figures 2 and 3(a)). The Pα emitting region seems to

be surrounded by the orbiting molecular gas. This prob-

ably suggests that the inner part of the Balloon’s orbit

has been ionized. The ionization may be attributed to

the dissociative shock caused by the sudden accelera-

tion of molecular gas by the strong gravitational force.

Hence, the detection of the Pα emission may also sup-

port for the presence of the IMBH, although a stellar

emission as the origin of the Pα cannot be ruled out.

Several observational studies have reported inter-

actions of outflows and/or intense radiation from

non-nuclear black holes with their ambient gas (e.g.

Soria et al. 2014; Tetarenko et al. 2018). For instance,

a persistent jet from Cygnus X-1 has been sug-

gested to create a shell-like structure of ionized gas

(Gallo et al. 2005; Russell et al. 2007). In addition,

the elongated ionized gas feature associated with M51

ULX-1 has been considered as evidence for an interac-

tion with a black hole jet (Urquhart et al. 2018). The

Pα emission feature we noticed could indicate a past

Indication of Another IMBH in the Galactic Center 7

activity of the IMBH although its spatial distribution

shows neither shell-like nor elongated morphology. Fur-

ther investigations of the kinematics and physical condi-

tions of the ionized gas would provide more convincing

evidence for the presence of an IMBH.

Several globular clusters and dwarf galaxies have been

suggested to harbor massive black holes with masses

lesser than 106 M as their nuclei (Reines et al. 2013;

Baldassare et al. 2015; Kızıltan et al. 2017). Recently,

it has been indicated that there is likely to be a

4 × 104 M IMBH in the center of ω Centauri, which

is the most luminous globular cluster in our Galaxy

(Baumgardt & Hilker 2018). The IMBH in HCN–

0.009–0.044 could be a remnant of such a globular

cluster. The parent cluster have possibly already been

dissolved before falling into the Galactic center.

The detected rotational motion of HCN–0.009–0.044

strongly indicates the presence of a dark massive object,

which is probably an IMBH. An emission counterpart

of the IMBH could be identified through future multi-

wavelength observations. Similar to the simulations con-

ducted for HVCC CO–0.40–0.22 (Ballone et al. 2018),

the hydrodynamical simulations of the molecular clouds

would more accurately restrict the orbital geome-

tries and IMBH mass. Although only ∼ 60 black

holes have been identified in our Galaxy to date

(Corral-Santana et al. 2016), the total number of black

holes in our Galaxy is theoretically estimated as ∼ 108–

109 (Agol & Kamionkowski 2002; Caputo et al. 2017).

Additionally, at least 104 black hole X-ray binaries

are observationally inferred to lurk in our Galaxy

(Tetarenko et al. 2016). These suggest that almost

all black holes are inactive with low accretion rates.

Regardless of the activity status of a black hole, its

strong gravity can disturb the ambient gas. As shown,

high-resolution observations of compact high-velocity

gas features have the potential to increase the number

of candidates for non-luminous black holes, providing a

new perspective to search for the missing black holes.

This paper makes use of the following ALMA data:

ADS/JAO.ALMA#2017.1.01557.S. ALMA is a part-

nership of ESO (representing its member states),

NSF (USA) and NINS (Japan), together with NRC

(Canada), MOST and ASIAA (Taiwan), and KASI

(Republic of Korea), in cooperation with the Republic

of Chile. The Joint ALMA Observatory is operated

by ESO, AUI/NRAO and NAOJ. We are grateful to

the ALMA staff for conducting the observations and

providing qualified data. We are also thankful to the

anonymous referee for helpful comments and suggestions

that improved this paper.

Facility: ALMA

Software: CASA (version 5.1.2)

REFERENCES

Agol, E. & Kamionkowski, M. 2002, MNRAS, 334, 553

Baldassare, V. F., Reines, A. E., Gallo, E. & Greene, J. E.

2015, ApJL, 809, L14

Ballone, A., Mapelli, M., & Pasquato, M. 2018, MNRAS,

480, 4684

Baumgardt, H. & Hilker, M. 2018, MNRAS, 478, 1520

Caputo, D. P., de Vries, N., Patruno, A. & Portegies Zwart,

S. 2017, MNRAS, 468, 4000

Corral-Santana, J. M., Casares, J., Munoz-Darias, T., et al.

2016, A&A, 587, 61

Djorgovski, S. & King, I. R. 1984, ApJL, 277, L49

Done, C. 2010, arXiv:1008.2287

Dong, H., Wang, Q. D., Cotera, A., et al. 2011, MNRAS,

417, 114

Ebisawa, K., Zycki, P., Kubota, A., Mizuno, T. & Watarai,

K.-y. 2003, ApJ, 597, 780

Ebisuzaki, T., Makino, J., Tsuru, T. G., et al. 2001, ApJL,

562, L19

Falcke, H., Kording, E., & Markoff, S. 2004, A&A, 414, 895

Farrell, S. A., Webb, N. A., Barret, D., Godet, O. &

Rodrigues, J. M. 2009, Natur, 460, 73

Fukui, Y., Sano, H., Sato, J., et al. 2012, ApJ, 746, 82

Gallo E., Fender R. P., Kaiser C., et al. 2005, Natur, 436,

819

Kızıltan, B., Baumgardt, H. & Loeb, A. 2017, Natur, 542,

203

McGary, R. S., Coil, A. L., & Ho, P. T. P. 2001, ApJ, 559,

326

Merloni, A., Heinz, S., & di Matteo, T. 2003, MNRAS, 345,

1057

Mezcua, M. 2017, IJMPD, 26, 1730021

Muno, M. P., Bauer, F. E., Baganoff, F. K., et al. 2009,

ApJS, 181, 110

Nomura, M., Oka, T., Yamada, M., et al. 2018, ApJ, 859, 29

Oka, T., Hasegawa, T., Sato, F., Tsuboi, M., & Miyazaki,

A. 1998, ApJS, 118, 455

Oka, T., Hasegawa, T., White, G. J., et al. 2008, PASJ, 60,

429

Oka, T., Nagai, M., Kamegai, K., & Tanaka, K. 2011, ApJ,

732, 120

Oka, T., Mizuno, R., Miura, K. & Takekawa, S. 2016,

ApJL, 816, L7

8 Takekawa et al.

Oka, T., Tsujimoto, S., Iwata, Y., Nomura, M. &

Takekawa, S. 2017, NatAs, 1, 709

Oka, T., White, G. J., Hasegawa, T., et al. 1999, ApJ, 515,

249

Parsons, H., Dempsey, J. T., Thomas, H. S., et al. 2018,

ApJS, 234, 22

Plotkin R. M., Markoff S., Kelly B. C., Kording E.,

Anderson S. F. 2012, MNRAS, 419, 267

Ravi, V., Vedantham, H. & Phinney, E. S. 2018 MNRAS,

478, L72

Reines, A. E., Greene, J. E. & Geha, M. 2013, ApJ, 775,

116

Russell D. M., Fender R. P., Gallo E., & Kaiser C. R. 2007,

MNRAS, 376, 1341

Schodel, R., Eckart, A., Iserlohe, C., Genzel, R. & Ott, T.

2005, ApJL, 625, L111

Sofue, Y. 2006, PASJ, 58, 335

Soria R., Long K. S., Blair W. P., et al. 2014, Science, 343,

1330

Strader, J., Chomiuk, L., Maccarone, T. J., et al. 2012,

ApJL, 750, L27

Takekawa, S., Oka, T., Iwata, Y., Tokuyama, S. & Nomura,

M. 2017, ApJL, 843, L11

Tanaka, K. 2018, ApJ, 859, 86

Tanaka, K., Nagai, M., Kamegai, K., Iino, T., & Sakai, T.

2018, ApJS, 236, 40

Tanaka, K., Nagai, M., Kamegai, K., & Oka, T. 2015, ApJ,

806, 130

Tanaka, K., Oka, T., Nagai, M., & Kamegai, K. 2009,

PASJ, 61, 461

Tetarenko, B. E., Bahramian, A., Arnason, R. M., et al.

2016, ApJ, 825, 10

Tetarenko, A. J., Freeman,P., Rosolowsky, E.W.,

Miller-Jones, J. C. A., & Sivakoff, G.R. 2018, MNRAS,

475, 448

Torii, K., Hattori, Y., Hasegawa, K., et al. 2017, ApJ, 835,

142

Tsuboi, M., Kitamura, Y., Tsutsumi, T., et al. 2017, ApJL,

850, L5

Urquhart, R., Soria, R., Johnston, H. M., et al. 2018,

MNRAS, 475, 3561

Yalinewich, A., & Beniamini, P. 2018, A&A, 612, L9

Zhao, J.-H., Morris, M. R., Goss, W. M., & An, T. 2009,

ApJ, 699, 186

Zhao, J.-H., Morris, M. R., & Goss, W. M. 2013, ApJ, 777,

146

Zhao, J.-H., Morris, M. R., & Goss, W. M. 2016, ApJ, 817,

171